Pilotage d'algorithmes pour la reconnaissance en ligne d'arythmies cardiaques

François Portet 1, 2
1 DREAM - Diagnosing, Recommending Actions and Modelling
Inria Rennes – Bretagne Atlantique , IRISA-D7 - GESTION DES DONNÉES ET DE LA CONNAISSANCE
Abstract : The aim of this work is the realization of the cardiac intelligent monitoring system IP-Calicot which is able, thanks to a piloting algorithms module, to use the current context information to modify dynamically its data processing chain to obtain a reliable medical diagnosis even in noisy situations. From an electrocardiogram (ECG), the system extracts on line, by signal processing, information which are used to establish a cardiac arrhythmia diagnosis modeled by a temporal network (chronicle). Using the current context, composed of the line noise and the medical diagnosis, the piloting module acts dynamically at three levels: it selects and tunes signal processing algorithms, it chooses the elements to be extracted from the ECG, which is thus described in a more or less precise language, and selects the description language to be used to establish the diagnosis on line. The pilot is represented by an expert system which acts on the data processing chain thanks to piloting rules acquired by expertise and deduced from statistical studies. The system has been validated on noisy ECG typical of clinical situations. The results show the interest and the feasibility of the proposed approach.
Complete list of metadatas

Cited literature [149 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-00011942
Contributor : François Portet <>
Submitted on : Wednesday, March 15, 2006 - 10:57:32 AM
Last modification on : Thursday, November 15, 2018 - 11:57:04 AM
Long-term archiving on : Monday, September 20, 2010 - 2:22:18 PM

Identifiers

  • HAL Id : tel-00011942, version 2

Citation

François Portet. Pilotage d'algorithmes pour la reconnaissance en ligne d'arythmies cardiaques. Intelligence artificielle [cs.AI]. Université Rennes 1, 2005. Français. ⟨tel-00011942v2⟩

Share

Metrics

Record views

1096

Files downloads

1052